Antarctic Science 14 (4): 395–411 (2002) © Antarctic Science Ltd Printed in the UK DOI: 10.1017/S0954102002000196 The sedimentary geology, palaeoenvironments and ichnocoenoses of the Lower Devonian Horlick Formation, Ohio Range, Antarctica MARGARET A. BRADSHAW1, JANE NEWMAN2 and JONATHON C. AITCHISON3 1Department of Geological Sciences, University of Canterbury, Private Bag 4800, Christchurch, New Zealand 2Newman Energy Research, 2 Rose Street, Christchurch 8002, New Zealand 3Department of Earth Sciences, University of Hong Kong, Hong Kong

Abstract: Six ichnocoenoses in the clastic Devonian Horlick Formation (max. 56 m) confirm the nearshore marine character of eight of the nine lithofacies present. A basal sand sheet overlies a weathered granitic land surface (Kukri Erosion Surface) on Cambro–Ordovician granitoids. The level nature of this surface and the way it cuts across weathering profiles, suggests that the surface had been modified by marine processes prior to deposition. The basal sand sheet (Cross-bedded Sand sheet Lithofacies) contains tidal bundles, and at its top, abundant Monocraterion (Monocraterion Ichnocoenosis). The second sand sheet (Pleurothyrella Lithofacies) is heavily burrowed and shows alternating periods of sedimentation, burrowing, and erosion below wave base as the sea deepened (Catenarichnus Ichnocoenosis). With increasing transgression, finer sediments were deposited (Laminated Mudstone and Feldspathic lithofacies) in an unstable pattern of coarse sandbars and finer troughs (Cruziana-Rusophycus and Arenicolites ichnocoenoses) crossed by active longshore marine channels (Poorly-sorted Lithofacies, Spirophyton Ichnoocoenosis). Short-lived but powerful storms produced thin shelly tempestites (Shell-bed Lithofacies), whereas sporadic, very thin phosphate rich beds (Phosphatic Lithofacies) may have resulted from marine transgressions across the basin. The deepest water is probably represented by sediments of the Spirifer Lithofacies (Rosselia Ichnocoenosis). The Schulthess Lithofacies is regarded as fluvial, deposited in the lower reaches of a river draining a land area that lay towards Marie Byrd Land. Channels in the basal sand sheet indicate movement to the south- west, but orientation became more variable higher in the sequence. Four new measured sections are figured. The relationship of the Ohio Range to the rest of Antarctica during the Devonian is suggested.

Received 6 March 2001, accepted 11 July 2002

Key words: Beacon Supergroup, sedimentology, trace fossils, Transantarctic Mountains

Introduction measured sections, was described in detail by McCartan & The Ohio Range is unique in that it is the only well Bradshaw (1987), who interpreted the environments of documented locality in Antarctica where there is a record of deposition as sub-tidal inner shelf to shoreline, with storms fossiliferous marine sediment of Devonian age. Elsewhere producing an unstable pattern of coarse sand bars and finer in the Transantarctic Mountains the Devonian is represented troughs. A moderate tidal range of between 1–3 m is likely. either by a hiatus or by very limited fossiliferous outcrops The shoreline was low-lying with a wave cut platform that (Ellsworth Mountains), or by nonfossiliferous sediment trimmed a deeply weathered and irregular land surface whose environmental interpretation is controversial, and underlain by granitic rocks. The level trimming of a whose age is uncertain. terrestrial landscape, followed by marine sediments, In the Ohio Range the Horlick Formation constitutes the strongly suggests marine modification of the Kukri Erosion lowest formation of the Beacon Supergroup and overlies the Surface in the Ohio Range. The source area for the Horlick level Kukri Erosion Surface cut across Cambro–Ordovician Formation lay northwards towards Marie Byrd Land, and a granitoids. The formation crops out along the northern face westerly longshore drift carried detritus away from river of the Ohio Range escarpment (Fig. 1), and preserved mouth deltas. The sequence was described in terms of nine thickness varies from 0 to 56 m due to the relief on an lithofacies (Table I), all but one of these being marine. Data overlying unconformity that is likely to have resulted from from four additional sections measured during a later field sub-glacial erosion during the late Palaeozoic. An Early visit are presented in Fig. 2. Devonian age for the Horlick Formation is indicated by an The shelly fauna of the Horlick Formation is dominated abundant shelly fauna (Boucot et al. 1963, Doumani et al. by the brachiopod Pleurothyrella antarctica Boucot et al. 1965). (1963) and a variety of large Malvinokaffric bivalves that The sedimentology of the formation, based on 15 confirm shallow nearshore sedimentation (Bradshaw &

395 396 MARGARET A. BRADSHAW et al.

Fig.1. Outcrop of the Horlick Formation (black) along the north-facing Ohio escarpment with the location of measured sections indicated.

Fig. 2. Detailed logs of Sections 16 (Darling Ridge), 17 (Darling Ridge), 18 (Discovery Ridge) and 19 (Canterbury Spur). ICHNOCOENOSES OF THE BASAL BEACON SUPERGROUP 397 sis Ichnocoenosis screet, short vertical Monocraterion Arenicolites None Ichnocoenosis Arenicolites & horizontal burrows , Ichnocoenosis. , . . Ichnocoenosis brachiopods, None Lingula, Catenarichnus , Obrimia Orbiculoidea ? Ichnocoenosis; , ; bryozoa. & Orbiculoidea , Modiomorpha, (broken in storm Nuculites ; rare fragments, fish plates Nuculites Nuculites , Ancryocrinus & , crinoid remains, bryozoa. Pleurothyrella Orbiculoidea, Obrimia, Spirophyton Pleurothyrella, Orbiculoidea,Pleurothyrella, Cruziana-Rusophycus Formation. Based on McCartan & Bradshaw (1987), Tanerhynchia Pleurothyrella, Orbiculoidea Pleurothyrella, Machaeracanthus, Burmeisteria Orbiculoidea, Lingula, Pleurothyrella; Burmeisteria Plectonotus, Prothyris,Australospirifer, Rosselia Obrimia Palaeosolen Pleurothyrella Obrimia thickness channelling & festoon bedding. scale hummocky cross stratification several times higher in Summary of lithological characters, depositional environments and faunal content the nine lithofacies present in Horlick (Lithofacies 8) parallel bedded; burrowing common. SchulthessLithofacies(Lithofacies 9) Medium to very coarse-grained, highly conglomerate; mudstone clasts; feldspathic sandstones to granule Fluvial channel 10 m Only on Schulthess Ridge. None None Table I. McCartan (1991), and trace fossils this paper. Lithofacies Characteristics Environment Max. Occurrence Body fossils fossils Trace Lithofacies sandstones; burrowing moderate subtidal sand bars. interbedded with Cross-bedded Coarse to very coarse-grained Marine; tidal to upper 10.5 m Basal sandsheet; thin Rare Sand sheetLithofacies(Lithofacies 1) cross-bedded sandstones; thin siltstones; thin basal conglomerate.Pleurothyrella Medium to coarse-grained transgression;unbarred shoreface, follows marine Middle shoreface; side of interbeds associated with Lithofacies 2 & 3 5 m finely comminuted fish bone; Forms second sandsheet; psilophyte plants. coastline. Ichnocoeno (Lithofacies 4) & parallel lamination; well sorted.Poorly-sortedLithofacies Coarse to very coarse-grained(Lithofacies 5) cross-bedded; shale clasts. sandstones; poorly-sorted; troughPhosphaticLithofacies Marine; deposited in long- shore channels or tidal delta(Lithofacies 6) Fine to medium-grained feldspathic 7.5 m phosphatic clasts common. sandstones; poorly sorted; channels near river mouths. Shell-bed Marine; lag deposits on Also interbedded withLithofacies(Lithofacies 7) Lithofacies 3 & 4. Medium to coarse-grained feldspathic 0.20 m carbonate cement. sandstones; poorly sorted; Rare shoreface sand-tongues. Marine; outer shorefaceSpirifer seaward flanks of outer Usually very thin; 0.25 m sequence. Fine to coarse-grained, muddy, Usually associated with to open shelf. Numerous molluscs, trilobites, repeated several times Marine; steeper parts of 3.5 m Only at top of sequence on Lithofacies 2 & 3. & bone; rare etc. (see Bradshaw & McCartan 1991). Ichnocoenosis Lithofacies feldspathic sandstones; poorly sorted; subtidal sandbars. Discovery Ridge. (Lithofacies 2) to intense.Laminated Interbedded mudstone & very fine- Marine; tidal influence; 2.5 m Follows second sandsheet; Lithofacies 3 & 7. MudstoneLithofacies(Lithofacies 3) grained sandstone; lenticular & wavy- bedded; interbedded storm deposits. flanks of sand-tongues &Feldspathic hollows between them.Lithofacies Micaceous fine-grained feldspathic sandstones; ripple lamination, small- Marine; top of sand-tongues; above wave-base. 2.5 m repeated several times Intimately associated with higher in sequence. deposits); rare Rare trilobite fragments, psilophyte plants. Lithofacies 3; repeated Di 398 MARGARET A. BRADSHAW et al.

Fig. 4. Lower part of Section 8 on Darling Ridge. The Kukri Erosion Surface below the first sand sheet (1–2 m thick) is clearly visible and truncates a weathered zone of variable thickness in the basement granitoid

Discovery Ridge and Treves Butte, the Kukri Erosion Surface truncates very large, altered basic inclusions in the basement granitoids (Fig. 3), but at most other localities the basement is homogenous quartz monzonite. Weathering of basement below the erosion surface is striking in some Fig. 3. Folded vein within a very large basic inclusion in basement sections (e.g. Section 8, Darling Ridge), but the depth of quartz monzonite truncated by the pre-Horlick erosion surface. weathered rock (0–10 m) varies greatly with location and The erosion surface is overlain by a coarse, poorly-sorted basal breccio-conglomerate of angular vein-quartz, feldspar and lithic appears to be a result of pre-Horlick trimming of an older, clasts. The black fragments above the erosion surface are undulating weathered land surface (Fig. 4). The level nature phosphatic. The top part of the basal bed and the bed following of the trimmed surface, as well as an Orbiculoidea shell show opposite sediment transport (herrirngbone structure) lying on the erosion surface at Echo Canyon, Lackey Ridge suggestive of tidal conditions. Arrow indicates hammer handle (Section 11), suggests that in the Ohio Range at least, the for scale. Kukri Erosion Surface achieved its final form as part of a wave-cut platform during the initial transgression of an Early Devonian sea (Bradshaw 1991). An extensively McCartan 1991). Fossils are usually confined to thin bands exhumed erosion surface is present on the north side of representing death assemblages that were probably West Spur Discovery Ridge (base Section 2), and a smaller concentrated by storms. Interbedded quartzose and feldspathic sandstones contain few shells but have a significant ichnofauna (Bradshaw et al. 1984, Bradshaw & McCartan 1991). This ichnofauna will be described fully in a later paper (Bradshaw unpublished data), and is only summarized here. This paper provides new information on the sedimentology of the Horlick Formation and its relationship to basement rocks, a review of its ichnofauna, a discussion of the palaeoenvironmental setting of the Ohio Range during the deposition of the Devonian, and a brief overview of the palaeogeography of the Transantarctic Mountains during the Devonian.

Relationship of Horlick Formation to basement Fig. 5. Wide channel, cut in basement quartz monzonite, infilled Like Devonian sediments elsewhere in the Transantarctic with sands of the Cross-bedded Sand sheet Lithofacies at the Mountains, the Horlick Formation overlies basement rocks base of Section 16, Darling Ridge. At this locality the basal sand with pronounced unconformity in the Ohio Range. On sheet is 3.5 m thick. Arrows points to basal contact. ICHNOCOENOSES OF THE BASAL BEACON SUPERGROUP 399

Fig. 6. Field drawing of serrated marine platform buried by sandstones of the Cross-bedded Sand sheet Lithofacies at the bottom of Section 18, east of the “Ice knife” on East Spur Discovery Ridge. exposure occurs near Echo Canyon (Lackey Ridge, base Section 11). Although the Kukri surface is level on a large scale, it sometimes shows a local relief of up to 2 m. Nowhere was the 20 m of relief seen that was reported by Long (1965). Post-Beacon Supergroup faulting occurs in the range, and Long’s relief may be simply a disparity of heights of the erosion surface at localities separated by hidden faults. The basal beds of the Horlick Formation are coarse, laterally discontinuous sandstones that infill depressions between low rounded domes (East Spur Discovery Ridge,

Fig. 8. Typical outcrop of the Horlick Formation showing the alternation of sandstone and mudstone lithologies. Sandstones of the Cross-bedded Sand sheet Lithofacies overlie basement quartz monzonite at the bottom of the photograph (black arrow). White arrow points to figure for scale. Type section for the Horlick Formation (Section 1) on East Spur Discovery Ridge.

Section 1), with the overlying continuous sand sheet often having a thin basal lag of angular quartz fragments. At other localities, very coarse, cross-bedded sandstones infill large channels down-cut into the basement (Darling Ridge, Section 16) and are followed by the main sand sheet (Fig. 5). At the base of Section 18 on the East Spur of a. Discovery Ridge, the Horlick Formation was deposited on a serrated marine platform, where coarse gravel accumulated in eroded troughs before the relief was buried by shoreward dipping tabular cross-beds (Fig. 6). On Lackey Ridge between Echo Canyon (Section 11) and Thumb Promontory (Section 12) two very large granitic, dome-like bodies (horizontal diameter 7 m and 5 m) are b. buried by the lowest Horlick sediments (Fig. 7). Whether the domes are attached to, or simply rest on, basement is not Fig. 7a. Rounded domes of quartz monzonite buried by sandstones clear in outcrop, but they appear to be residual giant domes of the Cross-bedded Sand sheet Lithofacies at the base of the of quartz-monzonite identical to the underlying basement. Horlick Formation on Lackey Ridge between Sections 11 and The domes show an 8 cm weathering rind, and the sediment 12. View looking south, uphill. Hammer for scale circled. Arrow infilling the space between them is very coarse-grained and x points to very coarse-grained, poorly-sorted feldspathic sandstone in between the domes. Arrow y indicates westward poorly-sorted sandstone rich in plagioclase feldspar. The dipping cross-sets of sandwave burying the domes. The domes are likely to represent the harder remnants of sediment here is well-sorted quartzarenite. b. Detailed field terrestrially exfoliated granitoid masses that were denuded drawing of contact of Horlick Formation sediment (left) against by wave-action during an Early Devonian transgression, quartz monzonite dome (right). and were eventually buried by the first sand sheet. Similar 400 MARGARET A. BRADSHAW et al.

Fig. 9. Coarse sandstones of the Cross-bedded Sand sheet Lithofacies in the basal sand sheet of Section 11 on Lackey Ridge. Bracket at right indicates possible tidal bundle. Arrow points to hummock-like drape. Scale in cm. domes may be the source of very large, rounded granitic boulders that occur in the overlying glacigenic Buckeye Formation not far above its base. The boulders were probably recycled from areas where the Horlick Formation was being actively removed during the Late Fig. 10. Tabular sandstone cross-set (Cross-bedded Sand sheet Carboniferous–Permian glaciation. Lithofacies), with fining up laminae and rippled bottomsets, burying a thin lag of phosphatic pebbles (not visible) that rests directly on phosphatized basement granite. Base Section 11, Horlick Formation: sediments, trace fossils, interpretation Echo Canyon, Lackey Ridge. Hammer for scale. The Horlick Formation comprises an alternating sequence of medium to coarse-grained sandstones and laminated however, the basal conglomerate is absent, and very coarse- mudstones, siltstones and fine sandstones (Fig. 8). The grained quartzose sandstone rests directly on basement (e.g. sequence was subdivided into nine numbered lithofacies Trilobite Promenade, Lackey Ridge, Section 14). At other (Bradshaw & McCartan 1983), which have been changed to localities, such as Section 16 (end Darling Ridge), the initial names in this paper and summarized in Table I. deposits are very coarse-grained to granule feldspathic Body fossils are common at certain horizons (see sandstones, and quartzose sandstones tend to be finer- Bradshaw & McCartan 1983), but lithologies in which they grained and confined to very thin interbeds. are scarce or absent, principally the sandstones, possess an The lower beds are generally cross-bedded, sometimes abundance of trace fossils. Six ichnocoenoses can be showing tidal bundles (Fig. 9, bracket), with thin siltstone recognized and are described below with the lithofacies in horizons. Some cross-sets possess hummock-like sandstone which they most commonly occur. drapes (Fig. 9, arrow). Near the base of Section 11 (Echo Canyon, Lackey Ridge), an extensive 1.70 m thick tabular cross-set with fining-up laminae and rippled bottom-sets Cross-bedded Sand sheet Lithofacies buries a channel lag of phosphatic pebbles that rest directly This lithofacies (Lithofacies 1 of McCartan & Bradshaw on phosphatized basement granite (Fig. 10). Ripples 1987) comprises coarse to very coarse-grained, cross- indicate a south-west flowing current. The cross-set bedded quartzose or feldspathic sandstones in a sand sheet laterally truncates very coarse-grained sandstones that is found at the base of most sections, and is repeated interbedded with lenticular bedded silty shales that pass higher in the sequence only as very thin interbeds associated southwards into wavy and flaser bedded sediments. The with sediments of the Pleurothyrella and Laminated outcrop pattern suggests a storm-wave dominated area Mudstone lithofacies. passing laterally into a large sandwave that was migrating In places, a thin basal conglomerate (up to 25 cm thick) is south-westwards along a channel bottom. present, consisting largely of fine to medium pebble-size, This lithofacies is sparsely fossiliferous (see Table I), and angular vein-quartz clasts, but locally including lithic, trace fossils are also scarce, largely consisting of the vertical phosphatic and feldspar clasts (Fig. 3). In some sections, burrows Monocraterion and Skolithos of the Monocraterion ICHNOCOENOSES OF THE BASAL BEACON SUPERGROUP 401

Fig. 11. Diagram to show component trace fossils of the six ichnocoenoses in the Horlick Formation (not drawn to scale).

Ichnocoenosis (Fig. 11a). The ichnocoenosis is best developed in horizontal and ripple-laminated fine to medium-grained sandstones at the top of the basal sand sheet where Monocraterion is common. Normally this is between 1–1.7 m above base (Sections 2, 11, 18), but where the basal sand sheet thickens, it occurs at 3.5 m (Section 16), or at 6 m and 10 m above base where the sand sheet is anomalously thick (Section 19). The bounding surfaces of beds containing this ichnocoenosis frequently possess numerous Monocraterion burrow openings (Fig. 12). Density of burrowing varies from relatively concentrated with 51 openings per m2 (Section 1, 12 m above base) to relatively sparse with 18 openings per m2 (Section 2, 1.5 m above base). The same bounding surfaces may contain scattered shells of Orbiculoidea, a shallow water Fig. 12. Bedding plane at the top of the lower sand sheet (Cross- inarticulate brachiopod. bedded Sand sheet Lithofacies) showing numerous burrow The appearance of abundant Monocraterion could be apertures of Monocraterion; Section 16, Darling Ridge. used as a marker horizon in the lower sand sheet, indicating Hammer for scale arrowed. 402 MARGARET A. BRADSHAW et al.

and 20 C/1 S respectively). As Woolfe (1993) has used this data to suggest nonmarine conditions for unfossiliferous Devonian sediment in southern Victoria Land, these suggestions have to be viewed with suspicion. The anomalous C/S ratios for the Ohio Range beds may be the result of mobilization and removal of sulphur.

Pleurothyrella Lithofacies (Lithofacies 2 of McCartan & Bradshaw 1987). Bioturbated, medium to coarse-grained sandstones containing Pleurothyrella, best developed in the second sand sheet. At a single locality on Darling Ridge (Section 9), Pleurothyrella Lithofacies sandstones rest directly on Fig. 13. Two Pleurothyrella Lithofacies horizons part way up the sequence in Section 11, Echo Canyon, Lackey Ridge. Two basement. episodes of burrowing are visible. Catenarichnus (c), The initial Pleurothyrella sand sheet is usually made up of ?Skolithos (s) and Diplocraterion (d) are recognisable. Scale four distinct sand units, each about 50 cm thick, separated in cm. by thin (c. 10 cm) laminated shales. In outcrop, these beds weather as four discreet, step-like exposures. The a change of bottom conditions as the transgression Pleurothyrella Lithofacies is repeated several times higher continued landwards. This period of Monocraterion in the sequence. formation is usually followed closely by the Pleurothyrella The sandstones are usually quartzose, sometimes poorly- Lithofacies. sorted, and often highly bioturbated. The structure of each Longitudinal sections of Monocraterian indicate unit varies considerably from locality to locality. Where prolonged colonization of the sandstones with upward bioturbation is intense, internal lamination is lost and the adjustment of the funnel-shaped openings to keep pace with bed has a hackly, yellow weathering appearance with sedimentation, so producing a series of stacked funnels. pockmarked bedding planes. However, less bioturbated Changes in grain size within the different laminae of the beds indicate amalgamation of sandstones with alternating burrow suggest that the animal producing Monocraterion cycles of burrowing and sedimentation, the latter sometimes was able to cope with sand grains of very different sizes. preceded by erosion (Fig. 13). Skolithos, on the other hand, is sparser and largely limited An anomalous, rounded, bryozoan-encrusted quartz to the basal beds of the sand sheet, consistent with its pebble, 4 cm in length, was found at the top of one reputation of preferring high energy environments and Pleurothyrella bed on Darling Ridge (Section 10) and shifting sands. suggests a prolonged break in sedimentation. A similar The Cross-bedded Sand sheet Lithofacies is interpreted as period of non-deposition is shown by the top of a bed on a wave-dominated, tidally influenced sequence of sand Lackey Ridge (Trilobite Promenade) near the base of deposits, with local short-lived areas of mud deposition in Section 14 which is current smoothed and studded with which plant material settled out. A sparse fauna of trilobites small phosphatic pebbles, fish bone and shells similar to the and brachiopod shells in both sandstones and mudstones Phosphatic Lithofacies (see below). Sediment banked suggests that all sediment was marine, and that the behind these clasts suggests transport from north-west to mudstones were not paralic as suggested by Long (1965). south-east, a similar current direction to that indicated by The near-shore region was crossed by deeper flow- parting lineation in sandstone interbeds in the overlying dominated channels. Ripple laminated finer sandstones Laminated Mudstone Lithofacies beds. To the south, the containing trace fossils suggest deeper water sedimentation smoothed surface is overlain by 90 cm of very coarse- and the establishment of Monocraterion colonies, with grained cross-stratified Orbiculoidea-bearing sandstone burrows keeping pace with sedimentation. that wedges out eastwards. This is interpreted as Arnot (1991) obtained unusually high C/S ratios for four interfingering of the Cross-bedded Sand sheet Lithofacies samples from very thin psilophyte-bearing shales in this with belts of Pleurothyrella and Laminated Mudstone lithofacies that suggest a freshwater environment, despite lithofacies further offshore. being enclosed by beds containing marine fossils and Shells are usually absent or rare in the strongly phosphatic material. The plant sample in the same bed as a bioturbated units, suggesting diagenetic solution, but are trilobite provided the highest C/S ratio for the Horlick more common where burrowing is less intense (see Table I). Formation (840 C:1 S, cf. Normal marine = 4 C:1 S), while Thick bryozoan encrustations were observed on some two other samples from a single thin psilophyte-bearing Pleurothyrella shells (Bradshaw & McCartan 1991 fig. 6). shale provided markedly dissimilar C/S ratios (470 C:1 S The Pleurothyrella Lithofacies contains the ICHNOCOENOSES OF THE BASAL BEACON SUPERGROUP 403

attempts by the animal to burrow more deeply as erosion occurred (c in Fig. 13). Narrow, poorly preserved, vertical to oblique burrows, tentatively identified as Skolithos (s in Fig. 13), postdate Catenarichnus. A vertical burrow 2 cm wide and 13 cm deep in the top part of one bed may have been formed by the burrowing bivalve Palaeosolen, external moulds of which have been recorded from the top of other beds (Bradshaw & McCartan 1991). The U-shaped burrow Diplocraterion parallelum, which has parallel arms and a meniscate structure, is also occasionally preserved (d in Fig. 13), especially in less bioturbated beds. At least one example showed retrusive (upward) behaviour, followed by protrusive (downward) Fig. 14. Cross-bedded sandstone unit of the Poorly-sorted behaviour, indicating fluctuating sedimentation rates. The Lithofacies, with shale (arrows) and fish bone clasts visible on the forsets, interbedded within the Laminated Mudstone vertical burrows Monocraterion, Cylindricum, and the Lithofacies. Section 1, East Spur Discovery Ridge. Brunton horizontal, back-filled burrow Olivellites, also occur. Compass for scale. The lithology and trace fossil content of the Pleurothyrella Lithofacies suggest that these beds were probably slowly deposited, with significant periods of non- Catenarichnus Ichnocoenosis (Fig. 11b), which is repeated deposition and erosion. The environment is likely to have several times above the second sand sheet in the same been below wave base but subject to moderate current lithofacies. Rosselia socialis, produced by a mud-filtering activity at times. The four layer construction of the second or deposit feeding organism, is particularly common, with sand sheet, with its very thin mudstone “spacers”, may funnels terminating at several different levels, indicating an reflect short-lived transgressions onto the land area during animal that was unable to adjust its burrow to repeated which time only mud accumulated on the sea floor beyond alternations of sedimentation and non-deposition. the surf zone. The transgressions were followed by longer Large arc-shaped Catenarichnus antarcticus (Bradshaw regressive periods of sand sedimentation below wave base 2002) is most obvious in the top of less bioturbated beds and the establishment of a prolific infauna. Rapid changes (Fig. 13), and the largest burrow observed was 37 cm in bottom current activity created alternating periods of between arc arms, and 11 cm deep. It is common to find sedimentation and erosion, which is reflected in the these burrows partially or largely removed by erosion burrowing behaviour of the infauna. Higher in the sequence before further deposition, and in some cases bedding planes this lithofacies probably developed on the sides of subtidal contain only the lowest portion of the burrow. Occasionally, sandbars. a deep and well preserved Catenarichnus burrow faithfully mirrors a shallower partially collapsed one, suggesting

Fig. 16. Sediments of the Laminated Mudstone Lithofacies grading up into ripple-laminated sandstones and parallel- Fig. 15. Thin interbed of Laminated Mudstone Lithofacies within laminated flaggy sandstones of the Feldspathic Lithofacies. Feldspathic Lithofacies sandstones, showing mud-draped A small channel has been eroded into the Laminated asymmetrical ripples with associated reactivation surfaces and Mudstone/Feldspathic Lithofacies couplet and is infilled with stoss side sedimentation that suggest tidal influence. Section 11, coarse sandstones of the Poorly-sorted Lithofacies. Section 3, Echo Canyon, Lackey Ridge. Scale in cm. West Spur Discovery Ridge. The depth of the channel is 30 cm. 404 MARGARET A. BRADSHAW et al.

Bradshaw 1987). Laminated Mudstone Lithofacies This lithofacies is intimately associated with the In this lithofacies (Lithofacies 3 of McCartan & Bradshaw Laminated Mudstone Lithofacies and the two may alternate 1987) lenticular and wavy-bedded mudstones are within a single unit. The Feldspathic Lithofacies is also interbedded with thinly laminated, very fine sandstones interbedded with erosive based sandstones of the Poorly- (Fig. 14). Mud-draped asymmetrical ripples with associated sorted Lithofacies, and was observed grading up into cross- reactivation surfaces and stoss side sedimentation at Echo bedded sandstones with shale clasts of the same lithofacies Canyon, Lackey Ridge (Section 11) suggest tidal influence (Darling Ridge, Section 16). (Fig. 15). The Laminated Mudstones are often interbedded Body fossils are rare (see Table I), but trace fossils are with thin horizons (e.g. 11 cm) of poorly sorted, very common, and two ichnocoenoses are present; the Cruziana- coarse, storm deposited sandstones containing fragmented Rusophycus Ichnocoenosis and the Arenicolites Pleurothyrella shells. Thicker interbeds with erosive bases Ichnocoenosis. The Cruziana–Rusophycus Ichnocoenosis are of cross-laminated sandstones of the Poorly-sorted (Fig. 11d) is confined to the more thickly bedded horizons, Lithofacies and have shale and bone clasts along foresets such as 13 m above base at Trilobite Promenade, Lackey (Fig. 14, arrow). Ridge (Section 14), although some of its elements are found A unit of Laminated Mudstone Lithofacies, 2–2.5 m in thinner sandstones in other sections. The ichnocoenosis thick, always follows the second sand sheet near the base of comprises Cruziana rhenana, Rusophycus, Imbrichnus, the succession, but the lithofacies is repeated several times Diplichnites gouldi, Diplichnites, Isopodichnus, higher in the sequence. Ancorichnus cf. capronus and Aulichnites. Body fossils are rare (see Table I) and trace fossils are Cruziana, Rusophycus and Diplichnites are all consistent generally not common, although more bioturbation was with surface trails produced by trilobite-like animals, observed at the eastern end of the range in this lithofacies although no skeletal material is known from these beds. (e.g. East Spur Discovery Ridge, Section 18), than at its However, disarticulated fragments of Burmeisteria western end. The ichnofauna is discreet, part of the (Digonus) antarcticus are particularly common in the Shell- Arenicolites Ichnocoenosis (Fig. 11c) which comprises bed Lithofacies, indicating that trilobites were present in the Arenicolites, Aulichnites, Palaeophycus and Olivellites. depositional basin. Some of these trilobites grew to a large Many of the short, irregular vertical and horizontal burrows size (width of the largest observed head shield was 18 cm), infilled with fine sandstone or siltstone are difficult to and although this size is consistent with the width of some identify. Rare horizontal burrows with horizontal meniscus of the Horlick Formation Diplichnites trackways, it is too have been identified as Teichichnus. large to have made the Cruziana and Rusophycus traces, The establishment of this lithofacies following the except as juveniles. deposition of the second sand sheet suggests that the sea Isopodichnus, a different trail, also occurs in this continued to deepen as the Early Devonian transgression lithofacies. The simplicity of appendage scratch marks progressed. The depositional area became affected by suggests that, although its producer was probably a small storms which produced a bar-and-trough bottom arthropod, it was not a trilobite. topography, with the laminated mudstones accumulating in The horizontal, back-packed burrow Ancorichnus cf. the troughs. The troughs experienced tidal currents as capronus is confined to fine-grained, rippled sandstones at shown by the nature of the ripples. Major storms swept Trilobite Promenade and in several other sections. It is coarser material out from the shore, and produced the thin, likely to have been made by an opportunistic animal poorly sorted sandstones with broken shell material that are following storm intervals. Opinion is divided as to whether interbedded with the mudstones. these animals were polychaetes or small arthropods. Imbrichnus, intimately associated with larger indeterminate Cubichnia, also has doubtful origins and The Feldspathic Lithofacies could have been created by either a mollusc (?gastropod; This lithofacies (Lithofacies 4 of McCartan & Bradshaw these are common in some beds) or arthropod (?crustacean). 1987) comprises fine-grained, well-sorted, commonly The poorly defined backfill of this prominent burrow, micaceous, feldspathic sandstones. Interference ripples are together with possible leg scratch marks and smooth, common, often with mud drapes. Thin, parallel laminated rounded marks at one end of resting depressions along the sandstones that show good parting lineation, and into which burrow, favour a crustacean animal. This animal was the ripple laminated sandstones frequently grade (Fig. 16), burrowing either immediately below the sediment/water are included in this lithofacies. Convoluted horizons up to interface or produced a surface trench in poorly 15 cm thick are occasionally present and ripple sets consolidated sediment as it methodically searched the sometimes show deformation. The ripple-laminated sediment for food, pushing material behind it as it went. sandstones contain the highest proportion of plagioclase to A single example of Aulichnites, possibly made by a potassium feldspar in the Horlick Formation ((McCartan & small crawling spired gastropod (e.g. Holopea), occurs ICHNOCOENOSES OF THE BASAL BEACON SUPERGROUP 405 alongside Rusophycus at Trilobite Promenade. Feldspathic Lithofacies, and also downcutting into The Arenicolites Ichnocoenosis is confined to horizontal Laminated Mudstone/Feldspathic lithofacies couplets (as in and rippled laminated beds in this lithofacies and Fig. 16). Current action was strong for the coarser beds and predominantly comprises the U-shaped, spreitenless burrow sedimentation was probably rapid. Arenicolites. These particular examples have deeper, more Body fossils are rare (see Table I) and usually confined to closely spaced, downward-diverging arms compared to the the better sorted, medium-grained sandstones, especially Arenicolites in other lithofacies, and the exhalent aperture where these are interbedded with Laminated Mudstone and lacks a sand collar. The U-shaped burrows indicate the Feldspathic lithofacies sediments. Trace fossils are also rare presence of worm-shaped animals, such as small and form the Spirophyton Ichnocoenosis (Fig. 11e), which holothurians, that were either suspension feeders, deposit comprises the burrows Spirophyton, Asterosoma, feeders, or a combination of both. Lanicoidichna, and “large footprints”. The assemblage is Both ichnofaunas suggest the passage of transient animals best seen in a thin coarse-grained sandstone 23 m above across the sediment surface and short-lived opportunistic base on East Spur Discovery Ridge (Section 1). shallow burrows after sedimentation. Deposition was The ichnocoenosis is characterized by large Spirophyton probably rapid. burrows infilled by mud from the overlying mudstone. The The Feldspathic Lithofacies represents well-sorted sand radiating burrow system Asterosoma in the same bed tongues that migrated across mud infilled hollows occasionally exhibit Spirophyton-like spreite ridges at the containing Laminated Mudstone Lithofacies sediment base of some of the blind radiating tunnels, suggesting a (McCartan & Bradshaw 1987). While the lower part of the common creator. Lanicoidichna, a ramifying burrow system sand tongues and adjacent troughs experienced tidal made up of both horizontal and vertical elements, is also influences, the top of the sand tongue was at, or above, common at this horizon and there are many instances where wave base. Sedimentation at wave base produced parts of this burrow system appear very like incomplete interference ripples, whereas sediments above this level Spirophyton. were deposited as low angle, thinly bedded, planar cross- It is possible that the three types of burrow reflect bedded laminae (Fig. 16). Parting lineation on the thin different behaviour by the same animal, with the foresets indicate a high flow regime, probably caused by complicated Spirophyton burrow produced when food was strong onshore wave action and shoreward movement of plentiful, dwelling burrows (Asterosoma) produced at other sediment, predominantly from the south-south-east to times, while exploratory tunnels excavated in search of food north-north-west. The sand bars were colonized by (Lanicoidichna) were created when food was short. burrowing animals, but many of these may have had to The “large footprints” found in a short single line on the relocate due to wave action. top of a poorly-sorted sandstone on Lackey Ridge, represent The high proportion of feldspar suggests that more fresh part of one side of a double row of superficial footprints. A material was accumulating in this lithofacies than elsewhere large arthropod animal (not a trilobite) is thought to have in the succession. It implies rapid transport and produced these prints, which have a completely different accumulation of material from a rugged granitic landscape pattern to those produced by trilobites. that was suffering little chemical weathering, possibly due The paucity of both trace fossils and body fossils in this to low temperatures. In other lithofacies, the feldspar was lithofacies, suggests that the environment was not amenable either mechanically worn down, or diagenetically reduced to life, probably because rapid sedimentation discouraged to clay in beds that were relatively slowly deposited, full of bottom faunas. Only in the more slowly deposited beds, or organic material and highly burrowed, such as the along the top of a deposited unit, do trace fossils become Pleurothyrella Lithofacies. Convoluted horizons and common. deformed ripples in the Feldspathic Lithofacies also suggest The erosional contact of this lithofacies with the rapid deposition. Feldspathic Lithofacies, and the occasional gradation between the two, suggests that the migrating feldspathic sandstone bars were bisected by active longshore marine or The Poorly-sorted Lithofacies tidal delta channels near river mouths, in which the Poorly- This lithofacies (Lithofacies 5 of McCartan & Bradshaw sorted Lithofacies sediment accumulated. 1987) includes trough cross-bedded, coarse to very coarse- grained, poorly-sorted sandstones with occasional shale Phosphatic Lithofacies clasts. Black phosphatic pebbles may be present above the erosional base of some sandstones. The Phosphatic Lithofacies (Lithofacies 6 of McCartan & This lithofacies can occur in units up to 7.5 m thick, or be Bradshaw 1987) comprises thin (1–20 cm), fine to medium- interbedded with horizons of Laminated Mudstone or grained, poorly-sorted feldspathic sandstones in which Feldspathic lithofacies (see Fig. 14). Sandstone beds were phosphatic clasts are common (phophatized mud pebbles, observed grading up into ripple laminated sediments of the fish bone and plates, trilobites and inarticulate 406 MARGARET A. BRADSHAW et al.

Spirifer Lithofacies The Spirifer Lithofacies (Lithofacies 8 of McCartan & Bradshaw 1987) is present only on Discovery Ridge having been probably removed elsewhere by Late Carboniferous– Permian glacial erosion. The lithofacies comprises parallel- bedded, poorly-sorted, muddy, fine to coarse-grained sandstones, with layers of subrounded quartz granules, or shell debris, and occasional cobbles of phosphatized sediment with bryozoan fossils (Fig. 17). Body fossils are common (see Table I) and include Australospirifer for the first time in the Horlick sequence. Crinoid holdfasts also occur, and some thick-shelled Fig. 17. Heavily burrowed, parallel-bedded, poorly-sorted molluscs have encrusting bryozoan colonies, which in some sandstones of the Spirifer Lithofacies at the top of Section 1, cases have become detached and buried separately (see East Spur Discovery Ridge. Hammer for scale. Bradshaw & McCartan 1991, fig. 5). The fine-grained sediment infill of articulated shells found in coarse brachiopods). This lithofacies is present in most sections sandstone (Bradshaw & McCartan 1991, fig. 4) suggest and though repeated several times, comprises only a small exhumation by storm waves from a siltstone host and proportion of the total succession. reburial in a coarser grained, higher energy environment. Body fossils are common (see Table I), but no trace Burrowing is abundant in this lithofacies with local fossils have been observed. destruction of primary lamination. Trace fossils comprise Most of the phosphate in the Horlick Formation is the Rosselia Ichnocoenosis (Fig. 11f), which is dominated concentrated into these beds. Phosphatization of sediment by long, well developed Rosselia socialis. Burrows extend just below the sediment-water interface can occur in downwards from many different horizons, suggesting epicontinental seas, seaways, and coastal embayments pauses in sedimentation, while erosion of the funnel during prolonged pauses in sedimentation, often in shallow entrance of some burrows, to leave only the sand infilled water (Cook 1984). Renewed wave action, sometimes lower tube, indicates a period of erosion after burrowing associated with marine transgression onto the land surface, and before renewed deposition. led to ripping up of the phosphatic sediment (and fossils) The base of some sandstones contain large Bergauria cf. and distribution of the debris, sediment and shell fragments, Langi. These are short vertical depressions that were as thin units. The phosphate rich beds are therefore thought excavated in the top of underlying mudstones later to represent short, renewed marine transgressions across the becoming infilled with coarse granular sandstone. The basin. Arenicolites Ichnocoenosis is present in thin horizons where Rosselia is absent. The Rosselia Ichnocoenosis is believed to have developed Shell-bed Lithofacies in a fully marine, probably middle shoreface environment, The Shell-bed Lithofacies (Lithofacies 7 of McCartan & supported by the presence of spiriferid brachiopods and Bradshaw 1987) comprises thin (25 cm or less), medium to crinoidal remains. Rosselia is likely to have been made by a coarse-grained, feldspathic, and calcareous sandstones in deposit or filter feeding worm-like animal that preferred which fossils are abundant (see Bradshaw & McCartan stable conditions. Bergauria may have had a sea anemone 1991 for details). The lithofacies is present in most sections, origin. often repeated several times, but overall is relatively rare. Arnot (1991) obtained high C/S ratios from a muddy The tops of some beds are crowded with current-oriented interbed within the Spirifer Lithofacies which, in his view, tentaculitids. indicated a brackish water environment of half normal Body fossils occur in dense accumulations and include salinity. This is difficult to accept when interbedded faunal mainly molluscs (bivalves and bellerophontids) evidence such as crinoid debris and spiriferid brachiopods brachiopods and trilobites. No trace fossils were observed. indicate fully marine offshore conditions. As mentioned in These beds are typical tempestites produced during major the Cross-bedded Sand sheet Lithofacies, anomalous C/S storms. Offshore shelly bottom faunas were probably ratios may be the result of leaching of sulphur and its exhumed by storm waves and their shells concentrated by concentration elsewhere. winnowing, with the lighter tentaculitid shell fraction being deposited last at the top of the bed as storm-energy waned. Schulthess Lithofacies The Schulthess Lithofacies (Lithofacies 9 of McCartan & ICHNOCOENOSES OF THE BASAL BEACON SUPERGROUP 407

Bradshaw 1987) comprises thick, cross-bedded, medium to thickness at other points, such as on Canterbury Spur very coarse-grained, highly feldspathic sandstone to granule (Section 19) where it is 10.5 m thick, and on Treves Butte. and pebble breccio-conglomerate, containing deformed Although Treves Butte could not be climbed because of mudstone clasts. Channel and festoon bedding are common. sheer cliffs and steep snow slopes, an American party, based This lithofacies is found only on west Schulthess Buttress in the Wisconsin Range, visited the site by helicopter in (Section 6) and sedimentary structures suggest that it was January 1965 and recorded 11 m of “barren” probably deposited in a fluvial channel setting (McCartan & (unfossiliferous) sandstone in the basal sand sheet below Bradshaw 1987). No body or trace fossils have been Pleurothyrella-rich sandstones at the bottom of a 21 m observed in this lithofacies. section (G.A. Doumani, personal communication 1981). Channel orientation in the basal sand sheet is predominantly north-east–south-west with movement of sediment towards Sedimentation summary the south-west. There is a recognisable sequential pattern to lithofacies in In all sections but Section 9 (Darling Ridge), the basal the lower part of most sections. Throughout the Ohio Range sand sheet is followed by a second sand sheet between 1 m the basal erosion surface is overlain by an almost and 2 m thick composed of burrowed Pleurothyrella continuous sand sheet composed largely of the marine Lithofacies sandstones. In Section 9 the Pleurothyrella Cross-bedded Sand sheet Lithofacies, but which at one Lithofacies rests directly on basement granite and suggests locality (Section 6) merges laterally into by the coarser onlap onto a local basement rise of several metres height. fluvial sandstones and breccio-conglomerates of the The second sand sheet is always followed by the Schulthess Lithofacies (west Schulthess Buttress). In Laminated mudstone Lithofacies, which is closely Section 9 on Darling Ridge the basal sand sheet is absent associated (both laterally and vertically) with moderately and at the extreme end of Lackey Ridge, north of Section 13 well-sorted, fine sandstones of the Feldspathic Lithofacies. (Panorama Point), it is very thin (80 cm). As well as at west In each section the lowest unit of the Laminated Mudstone Schulthess Buttress (10 m), the sand sheet increases in Lithofacies is relatively thick (e.g. 2.5 m, Section 17), but higher in the sequence it occurs as thinner interbedded units. The upper part of the Horlick Formation tends to be more variable. Coarse-grained sandstones of the Poorly-sorted Lithofacies are found interbedded with repetitions of the Laminated Mudstone and Feldspathic lithofacies, while thin units of Phosphatic and Shell Bed lithofacies are found scattered throughout all but the top of the sequence. An attempt was made to correlate the shell-bed horizons, and four main beds appear to be present; one 6–7 m above base, a second at 10–13 m, a third at 15–19 m and a fourth at 36–38 m above base. However, correlation was difficult because the thickness of many sections has been reduced by pre-Permian erosion, and at only five sections (1, 2, 3, 10 and 18) does the succession exceed 30 m. It was noted that the thin phosphatic pebble horizons, possibly marking transgressive events, occur at roughly similar heights above base as do the shell-bed horizons (e.g. 5–6 m above base, 9–11 m, 12–15 m, 17–18 m), suggesting that the shell beds may have developed not long after significant transgressive events. The Spirifer Lithofacies is confined to the top of the thickest sequence (East Spur Discovery Ridge sections 1, 2 & 18). Channel orientation and sediment movement seem more variable in the higher sandstones than in the initial sand sheet. Parting lineation, found principally in low angle planar-bedded sandstones of the Feldspathic Lithofacies, has a predominant north-west-south-east direction, with a secondary north-east-south-west lineation, and a less Fig. 18. Aggregate percentages of different lithofacies in four common north-south lineation. sections along the Ohio Range escarpment. For detailed logs of The proportion of different lithofacies varies throughout sections 16–19 see Fig. 2. For all other sections see McCartan & the range, and lateral changes are rapid. However, there Bradshaw (1987). seem to be broad differences in the percentages of 408 MARGARET A. BRADSHAW et al. lithofacies present between the western end of the Ohio Palaeoenvironmental summary escarpment (Darling Ridge, Section 16) and the eastern end (“Ice knife” on East Spur Discovery Ridge, Section 18) The lowest sand sheet was produced by active (Fig. 18). For example, at the western end 64% of the total sedimentation associated with a marine transgression. It section is composed of the finer elements of the formation, was followed by the second sand sheet representing represented by a combination of the Laminated Mudstone periodic but relatively uniform deposition over a broad area and Feldspathic lithofacies, compared with only 30% at the in an epeiric sea, with each phase followed by profuse eastern end. It was noted that burrowing is more obvious in bioturbation, and exhumation and disarticulation of shallow the Laminated Mudstone Lithofacies at the eastern end of water brachiopods (Figs 13). The lower two sand sheets the range compared with the west. The Cross-bedded preceded the development of a bar-and-trough bottom Sandstone Lithofacies is better represented in the west (20% topography represented by the Laminated Mudstone and W cf. 3% E), but the Pleurothyrella Lithofacies (7% W cf. Feldspathic lithofacies, where gradations between the two 35% E), Poorly-sorted Lithofacies (7% W cf. 20% E), lithofacies are common. Phosphatic Lithofacies (2% W cf. 5% E) and the Shell-bed The occasional gradation of the Feldspathic Lithofacies Lithofacies (0% W of the sequence cf. 7% E) are reduced. into the Poorly-sorted sandstone Lithofacies, without the Between Darling and Discovery ridges, on Schulthess more usual interdigitation, or an erosional contact, suggests Buttress W (Section 6), the entire 10 m sequence is that the migrating bars were bisected by active longshore comprised of uniform sandstones of the Schulthess marine channels, or tidal delta channels near river mouths. Lithofacies, probably deposited in the lower reaches of a The thin beds of the Phosphatic Lithofacies are likely to river mouth. More typical Horlick lithologies lie 4 km to the represent short, renewed marine transgression onto the east (Section 5) and 7 km to the west (Section 17) into adjacent land area, the effects of which were felt across the which the Schulthess Lithofacies must merge laterally over entire basin. these distances. The most easterly measured section in the The Shell-bed Lithofacies represents typical tempestites, range (Canterbury Spur, Section 19, see Fig. 2) is 10.5 m with the shells of bottom faunas becoming concentrated by thick and is also composed entirely of sandstone, in this wave action. The shell beds represent major storm events case comparable to the Cross-bedded Sand sheet and that were widespread across the depositional basin. Poorly-sorted lithofacies. Although the section contains no The Schulthess Lithofacies is recognized by its body fossils, marine trace fossils (e.g. Monocraterion, coarseness and lack of body and trace fossils. It occurs Asterosoma) are present at several horizons. This section clearly at only one locality, but may be present to a lesser may also have been deposited close to a river mouth delta degree at the base of the sequence on Canterbury Spur, and that was interfingering west with the coarse sandbars and possibly also on Treves Butte. finer troughs represented by Section 18 on the east side of Discovery Ridge 3 km away (Fig. 2), and north-west with Antarctic Devonian palaeogeography the “barren” sandstone unit on Treves Butte. The unusually thick sandstone units in Section 18 between 7 & 10 m and Two facts prevent a comprehensive overall reconstruction 24 & 31 m tend to support this. of Antarctic sedimentary basins during the Devonian. A high proportion of basal sand sheet sandstone is quartz- Firstly, most localities do not show a conformable rich due to removal of feldspar in the basement regolith by passage upwards from the highest preserved Devonian beds weathering. McCartan & Bradshaw (1987) record less than into Carboniferous or younger beds. In most cases the 5% feldspar in some sections. Devonian is truncated by the Maya Unconformity below the Thickening of the basal sand sheet, infilling of channel Late Carboniferous–Permian glacial sequence, and an structures and an increase in the proportion of feldspar unknown amount of Devonian sediment may have been proximal to the Schulthess Lithofacies at Schulthess removed throughout the Transantarctic Mountains. The Buttress is consistent with accumulation close to a large exception is the Ellsworth Mountains where the Wyatt Earp river mouth that was discharging freshly eroded material Formation is considered to be Devonian in age, with a into the area, where it was rapidly deposited. Elsewhere, the maximum thickness of 300 m in the Sentinel Range. In the basal sandstones were probably accumulating in more Heritage Range, where the only Devonian fossils in the shallow water, closer inshore, where less feldspar survived. formation are found, the top of the sequence contains Longshore drift was westward and another river mouth to interbedded conglomerate similar to the overlying Late the east must have supplied sediment to the sections Carboniferous–Permian Whiteout Conglomerate Formation between Schulthess Buttress and Discovery Ridge (Spörli 1992). In the adjacent Sentinel Range, the upper part (McCartan & Bradshaw 1987). The thick basal sandstones of the Wyatt Earp Formation contains isolated pebbles that on Canterbury Spur and Treves Butte may mark the position are thought to have been ice rafted. A continuous section of this second river mouth. into Carboniferous rocks appears to be present in both ranges, although the Devonian–Carboniferous boundary ICHNOCOENOSES OF THE BASAL BEACON SUPERGROUP 409 cannot be identified. In places there is deformation of tempting to correlate the surface below this conglomerate sediment below the conglomerate and local erosional with the Kukri Erosion Surface. Storey et al. (1996), while contacts (Spörli 1992), but this could be intra- agreeing that the Dover Sandstone could be correlated with Carboniferous. part of the Beacon Supergroup, considered that the Secondly, where the lower part of the Taylor Group underlying 3000 m thick Neptune Group was older, and (Beacon Supergroup) rests unconformably on older rocks, likely to have been deposited while the Ross Orogeny was the beds are difficult to date at most localities because no taking place. They interpret most of the Neptune Group as a body fossils are present. The exceptions are the Ohio syntectonic alluvial fan complex, but saw the highest Range, where a shelly fauna (no graptolites, conodonts or formation, the Heiser Sandstone, as marine. As there is no ammonoids) suggests an Emsian (lower Devonian) age, and fossil control for the Neptune Group, the age of this marine the McMurdo Dry Valley region, where poorly preserved transgression, and how it might relate to the Ohio Range, is palynomorphs suggest an age little older than Emsian. unknown. Middle to Late Devonian fish fossils in southern Victoria Although there can be no question that the Ohio Range Land provide a reliable age for the upper part of the Taylor and Ellsworth Mountains sedimentary basins were linked Group below the glacial Maya Unconformity, but these beds during the early Devonian due to the similarity of their (Aztec Formation) exist only between the Mackay and faunas, it is unlikely that the Ellsworth-Ohio basin Darwin glaciers. communicated with the central Transantarctic Mountains The Devonian of the Ellsworth Mountains (Wyatt Earp basin. Isbell (1999) has listed evidence which suggests that Formation) occurs at the top of a very thick, apparently rather than the glacial dissection of a continuous sheet of conformable, Palaeozoic clastic sequence (Crashsite Taylor Group sediments from the Ohio Range to Victoria Group), which ranges from Late Cambrian (Shergold & Land, as previously thought, it is more likely that Taylor Webers 1992) to Devonian in age (Webers et al. 1992). A Group sedimentation occurred in separate intermontane or limited Devonian fauna was collected from a 128 m thick “successor” basins, separated by basement highs. This is section of Wyatt Earp Formation near Planck Point, supported by the onlap of Late Carboniferous–Permian Heritage Range (Webers et al. 1992), and provided the first sediments onto a high between the Amundsen and Ramsey definite age. The fossils, which occur in concretions in glaciers along the margins of the central Transantarctic sandstone, are identical to those found in the Ohio Range Mountains basin (Queen Maud High of Collinson 1990), (personal observation) and a similar Emsian age is likely. and by the fact that the glacial sequence tends to be thick Deposition occurred under shallow marine conditions. where the Taylor Group is thick, and thin where the latter is Spörli (1992) also noted trace fossils and “shredded plant absent. Isbell considers that the Taylor Group filled material”. depressions on the eroded Ross orogenic belt. This is borne The association of the Ohio Range fauna with that of the out in the Central Transantarctic Mountains basin where, highest formation of the Crashsite Group in the Ellsworth near the Starshot Glacier, a 5 m basal conglomerate is Mountains is interesting because the Ohio Range basement followed by 123 thick shale unit Castle Crags Formation) is a Ross Orogeny granitoid (?Cambrian) and the Ellsworth below the more characteristic Beacon quartzarenites basin had Cambrian sedimentation in a probable rift (Alexander Formation). Isbell suggests deposition in a environment (Curtis 2001). The two areas became depression on the Kukri erosion surface, possibly in a connected only when erosion had unroofed the lacustrine environment. In the Churchill Mountains south of Cambro–Ordovician Ross granitoids, and the sea the Byrd Glacier, sandstones and local conglomerates bury transgressed onto the roots of the Ross mountain chain to a paleokarst topography developed on deformed Cambrian deposit the Horlick formation in the Ohio Range, and the marble, with a relief of up to 700 m (Anderson 1979). Wyatt Earp Formation in the Ellsworth basin. The Devonian sequence of the southern Victoria Land and In the Devonian, the Pensacola Mountains would have Darwin Glacier region is generally accepted as forming in lain between the Ohio Range and the Ellsworth Mountains another intracratonic basin north-west of a ‘Ross’ basement before the Ellsworth-Whitmore block rotated from a high that lay south of the Byrd Glacier (Barrett 1981, 1991, position adjacent to South Africa (Grunow et al. 1991) Collinson et al. 1994, Woolfe & Barrett 1995). Up to during late Permian earth movements. How much of the 1500 m of predominantly arenaceous sediments accumulated Pensacola Mountains succession is Devonian is not clear. in this basin, and there is evidence of onlap of the lower Haplostigma from an isolated outcrop of sandstones and Taylor Group formations onto another basement high to the shales has been used to suggest that the highest formation north (Balham Valley High, Bradshaw 1981), with sediment below the glacial Gale Mudstone (Dover Sandstone) could transport from the Ross Sea side of the basin as well as from be Middle Devonian in age (see Bradshaw & Webers 1988 the East Antarctic Craton. More is known about the for review), although the range for this plant genus is broad. Devonian sedimentation of this region than for the Central The Dover Sandstone is 1200 m thick, and has a thin basal Transantarctic basin, and there is unresolved controversy conglomerate of rounded quartz and phosphate pebbles. It is over whether the lower part of the Taylor Group is coastal 410 MARGARET A. BRADSHAW et al. marine or fluvial. Trace fossils and sedimentary structures Summary and conclusions locally suggest a marine influence (Bradshaw 1981, The Horlick Formation accumulated adjacent to a tide and Bradshaw & Webers 1988), but elsewhere other features storm dominated coastline. A thin initial sand sheet was suggest a fluvial influence (Woolfe 1993), pointing to an deposited on a wave-trimmed platform, in places burying intimate mixing of the two environments in a coastal belt. wave-cleaned domes of exfoliated granite, and merging into The proliferation of trace fossils, including ichnogenera thick fluvial sediments near river mouths. A thin second normally associated with marine environments, and present sand sheet deposited beyond the wave zone became in the contemporaneous marine Horlick Formation intensely bioturbated during episodic deposition, after (Diplichnites, Skolithos, Arenicolites, Cruziana, which tidally influenced migratory sand shoals and muddy Rusophycus, Aulichnites, “large footprints”), suggest that hollows developed. Lithologies are repeated higher in the the possibility of periodic marine incursions should be taken succession, with short, powerful storms carrying a shelly seriously. Ichnofaunas in the Darwin and Cook Mountains, benthic fauna shoreward, and minor transgressive phases and their comparison with those of the Dry Valley region, producing thin, phosphate-rich horizons. will clarify this story (M.A. Bradshaw, unpublished data). The sandstones in particular were home to a variety of There is no dispute about the alluvial origin of the Middle to organisms, many of them arthropods. Freshwater and Late Devonian Aztec Siltstone Formation at the top of the brackish type C/S ratios obtained from horizons that contain Taylor Group (dated on fish fossils, Young 1988), which occasional marine fossils suggest that this technique should contains typical alluvial plain sediments, palaeosols and be used with caution and does not accurately identify root horizons (McPherson 1979). The Aztec Siltstone marine conditions in the Horlick Formation. Formation overlies the Beacon Heights Orthoquartzite, which occasionally contains Haplostigma, suggesting a middle Devonian age. There is southward thinning of both Acknowledgements formations in the Cook Mountains south of the Mulock Thanks to Molly Miller who made pertinent suggestions to Glacier (M.A. Bradshaw, unpublished data). an earlier draft of the manuscript and to John Isbell, Morag The southern Victoria Land basin may have extended well Hunter and Nigel Trewin for their thorough reviews of the to the east below the Ross Sea. Multichannel seismic paper. Thanks also to mountain guide Bill Atkinson for his profiles indicate 6–7 km of stratified sediments, likely to be companionship in the field. The preparation of this Beacon Supergroup and older Palaeozoic rocks, below manuscript has been significantly assisted by a grant to Cenozoic rocks in the Victoria Land basin (Cooper & Davey MAB from the TransAntarctic Association. 1985). The combined thickness of the Beacon Supergroup (Taylor and Victoria Groups) exposed on land is 2.5 km. In late 1999, the Cape Roberts Sea Floor Drilling Programme References on the western side of the Victoria basin terminated in ANDERSON, J.M. 1979. The geology of the Taylor Group, Beacon middle Taylor Group sediments (?Arena Sandstone) after Supergroup, Byrd Glacier area, Antarctica. New Zealand Antarctic passing through nearly 940 m of Cenozoic sediments (M.G. Record, 2, 6–11. Laird, personal communication 2001).This suggests that the ARNOT, M.J. 1991. C/S geochemistry of Beacon Supergroup rocks, from whole of the Victoria Group (Permian–Triassic) and the top southern Victoria Land and the Ohio Range, Transantarctic Mountains, Antarctica MSc. thesis, Victoria University, Wellington, 72 pp. part of the Taylor Group was removed before Cenozoic [Unpublished]. sedimentation commenced. BARRETT, P.J. 1981. History of the Ross Sea region during the deposition of Woolfe & Barrett (1995) inferred that the marine the Beacon Supergroup 400 to 180 million years ago. Journal of the palaeopacific margin in the Devonian lay at least 1000 km Royal Society of New Zealand, 11, 447–458. east of the southern Victoria Land Taylor Group basin, BARRETT, P.J. 1991. The Devonian to Jurassic Beacon Supergroup of the Transantarctic Mountains and correlatives in other parts of Antarctica. In based on the extent of the pre-Ross Swanson Group, and TINGEY, R.J., ed. The geology of Antarctica. Oxford: Clarendon Press, allowing for c. 33% extension during the Cenozoic. 120–152. However, the existence of Swanson Group in Marie Byrd BOUCOT, A.J., CASTER, K.E., IVES, D. & TALENT, J.A. 1963. Relationships Land does not indicate the position of the Early Devonian of a new Lower Devonian terebratuloid (Brachiopoda) from Antarctica. coastline. In New Zealand (at Reefton) Greenland Group Bulletins of American Paleontology, 46, 81–151. BRADSHAW, M.A. 1981. Paleoenvironmental interpretations and rocks, the probable equivalent of the Swanson Group, are systematics of Devonian trace fossils from the Taylor Group (lower unconformably overlain by thick Lower Devonian Beacon Supergroup), Antarctica. New Zealand Journal of Geology and nearshore sandstones, and shelf limestones and mudstones. Geophysics, 24, 615–652. It is possible that these marginal Gondwana seas extended BRADSHAW, M.A. 1991. The Devonian Pacific margin of Antarctica. In much further west and communicated with epicontinental THOMSON, M.R.A., CRAME, J.A. & THOMSON, J.W., eds. Geological evolution of Antarctica. Cambridge: Cambridge University Press, seas further inland to provide the marine influence shown 193–197. by the sediments of southern Victoria Land and their ichnofauna (Bradshaw 1999). ICHNOCOENOSES OF THE BASAL BEACON SUPERGROUP 411

BRADSHAW, M.A. 1999. Lower Devonian bivalves from the Reefton GRUNOW, A.M., KENT, D.V. & DALZIEL, I.W.D. 1991. New paleomagnetic Group, New Zealand. Association of Australasian Palaeontologists data from Thurston Island: implications for the tectonics of West Memoir, 20, 1–171. Antarctica and Weddell Sea opening. Journal of Geophysical Research, BRADSHAW, M.A. 2002. A new ichnogenus Caternarichnus from the 96, 17 935–17 954. Devonian of the Ohio Range, Antarctica. Antarctic Science, 15, ISBELL, J.L. 1999. The Kukri Erosion surface; a reassessment of its 422–424. relationship to rocks of the Beacon Supergroup in the central BRADSHAW, M.A. & MCCARTAN, L. 1983. The depositional environment of Transantarctic Mountains. Antarctic Science, 11, 228–238. the Lower Devonian Horlick Formation, Ohio Range. In OLIVER, R.L., LONG, W.E. 1965. Stratigraphy of the Ohio Range, Antarctica. Antarctic JAMES, P.R. & JAGO, J.B., eds. Antarctic earth science. Canberra: Research Series, 6, 71–116. Australian Academy of Science & Cambridge: Cambridge University MCCARTAN, L. & BRADSHAW, M.A. 1987. Petrology and sedimentology of Press, 238–241. the Horlick Formation (Lower Devonian), Ohio Range, Transantarctic BRADSHAW, M.A. & MCCARTAN, L. 1991. Palaeoecology and systematics Mountains. Bulletin US Geological Survey, 1780, 31 pp. of Lower Devonian bivalves from the Horlick Formation, Ohio Range, MCPHERSON, J.G. 1979. Calcrete (caliche) palaeosols in fluvial red beds of Antarctica. Alcheringa, 15, 1–42. the Aztec Siltstone (Upper Devonian), southern Victoria Land, BRADSHAW, M.A., NEWMAN, J. & AITCHISON, J.C. 1984. Preliminary Antarctica. Sedimentary Geology, 22, 267–285. geological results of the 1983–84 Ohio Range expedition. New Zealand SHERGOLD, J.H. & WEBERS, W.F. 1992. Late Dresbachian (Idamean) and Antarctic Record, 5, 1–17. other trilobite faunas from the Heritage Range, Ellsworth Mountains, BRADSHAW, M.A. & WEBERS, G. 1988. The Devonian rocks of Antarctica. West Antarctica. In WEBERS, G.F., CRADDOCK, C. & SPLETTSTOESSER, In MCMILLAN, N.J., EMBRY, A.F. & GLASS, D.J., eds. Devonian of the J.F., eds. Geology and paleontology of the Ellsworth Mountains, West world. Calgary: Canadian Society of Petroleum Geologists, Memoir 14, Antarctica. Geological Society of America Memoir, No. 170, 125–168. 783–795. SPÖRLI, K.B. 1992. Stratigraphy of the Crashsite Group. In WEBERS, G.F., COLLINSON, J.W. 1990. Depositional setting of Late Carboniferous to CRADDOCK, C. & SPLETTSTOESSER, J.F., eds. Geology and Paleontology Triassic biota in the Transantarctic Basin. In TAYLOR, T.N. & TAYLOR, of the Ellsworth Mountains, West Antarctica. Geological Society of E.L., eds. Antarctic palaeobiology: its role in the reconstruction of America Memoir, No. 170, 21–35. Gondwana. New York: Springer, 1–14. STOREY, B.C., MACDONALD, D.I.M., DALZIEL, I.W.D., ISBELL, J.L. & COLLINSON, J.W., ISBELL, J.L., ELLIOT, D.H., MILLER, M.F. & MILLER, MILLAR, I. 1996. Early Paleozoic sedimentation, magmatism and J.M.G. 1994. Permian–Triassic Transantarctic basin. In VEEVERS, J.J. & deformation in the Pensacola Mountains, Antarctica: the significance of POWELL, C.MCA., eds. Permian–Triassic Pangean Basins and Foldbelts the Ross Orogeny. Geological Society of America Bulletin, 108, along the Panthalassan Margin of Gondwanaland. Geological Society 685–707. of America Memoir, No. 184, 173–222. WEBERS, W.F., GLENISTER, B., POJETA JR, J. & YOUNG, G. 1992. Devonian COOK, P.J. 1984. Spatial and temporal controls on the formation of fossils from the Ellsworth Mountains, West Antarctica. In WEBERS, G.F., phosphate deposits - a review. In NRIAGU, J.O. & MOORE, P.B., eds. CRADDOCK, C. & SPLETTSTOESSER, J.F., eds. Geology and paleontology Phosphate minerals. Berlin: Springer, 242–274. of the Ellsworth Mountains, West Antarctica. Geological Society of COOPER, A.K. & DAVEY, F.J. 1985. Episodic rifting of Phanerozoic rocks in America Memoir, No. 170, 269–278. the Victoria Land Basin, western Ross Sea, Antarctica. Science, 229, WOOLFE, K.J. 1993. Devonian depositional environments in the Darwin 1085–1087. Mountains: marine or non-marine? Antarctic Science, 5, 211–220. CURTIS, M.L. 2001. Tectonic history of the Ellsworth Mountains, West WOOLFE, K.J. & BARRETT, P.J. 1995. Constraining the Devonian to Triassic Antarctica: reconciling a Gondwana enigma. Geological Society of Tectonic evolution of the Ross Sea Sector. Terra Antarctica, 2, 7–21. America Bulletin, 113, 939–958. YOUNG, G.C. 1988. Antiarch (placoderm fishes) from the Devonian Aztec DOUMANI, G.A., BOARDMAN, R.S., ROWELL, A.J., BOUCOT, A.J., JOHNSON, Siltstone, southern Victoria Land, Antarctica. Palaeontographica, 202A, J.G., MCALESTER, A.L., SAUL, J., FISHER, D.W. & MILES, R.S. 1965. 1–125. Lower Devonian fauna of the Horlick Formation, Ohio Range, Antarctica. Antarctic Research Series, 6, 241–281.